Gas Cars Just Dipped Below 50% Globally — but the US P&L Has Already Moved, and Not From the Milestone

Per Nikkei Asia and Mobility Global, petrol-only vehicles fell to 49% of global new-vehicle sales in H1 2026, down from 52% a year earlier and 73% in 2021. Absolute petrol sales were 20.25 million units, down 10% year-on-year — the first sub-50% half on record. China led the retreat at -26%, Europe at -13%. Higher oil prices \"triggered by conflict in the Middle East pushed more consumers toward electric vehicles,\" Nikkei reported, citing data compiled by Mobility Global.

The US did not cross the line. Through Q2 2026, electrified vehicles (BEV + PHEV + HEV) were 26.9% of US retail new-vehicle sales per Urban Science — leaving ICE at ~73%. Battery-electric alone was 7.4%, down 13.9% YoY; plug-in hybrids were 1.4%, down 26.6% YoY. The entire growth in electrified share came from hybrids at 18.1%, up 26.2% YoY — and hybrids still sip gasoline, so they extend rather than end the ICE fuel cycle.

The P&L reaction that actually showed up in Q2 2026 belongs to the refiners. The Iran war drove Brent from roughly $65/bbl in early February to above $100/bbl by mid-September, and the WTI 3-2-1 crack spread climbed to about $59/bbl. Valero Energy reported Q2 revenue of $44.5 billion (17% above consensus) and a refining margin per barrel of $23.62, almost double $12.35 a year earlier. Refining-segment operating income reached $4.5 billion, against $1.3 billion in Q2 2025. Marathon Petroleum reported net income of $5.1 billion and an R&M margin of $36.33/bbl, more than double $17.58/bbl a year earlier, returning $2.8 billion to shareholders in the quarter.

Marathon Petroleum, Valero and HF Sinclair each returned 80%+ year-to-date through mid-August, against an S&P 500 up 11%, per CNBC. The windfall looks durable as long as Middle East tension keeps product inventories tight. The IEA's September 2026 Oil Market Report, however, cut 2026 global oil demand by 940 kb/d and now expects a 2.5 mb/d decline for the year — the demand-side picture that eventually closes the window. Refining margins are forward-pricing a supply shock today while the demand-curve is sloping down underneath them.

The ICE <50% headline obscures the more durable earnings stream in the replacement-parts business. Genuine Parts Company reported North America Automotive segment sales of $2.5 billion in Q2 2026, up 3.8% YoY, on a 2.6% increase in comp store sales. Total revenue was $6.5 billion (+6%); adjusted EPS of $2.15 was up 2.5%. The company reaffirmed its 2026 adjusted EPS guidance of $7.50–$8.00 and is pressing ahead with a planned separation of its industrial and automotive segments.

The economic logic is mechanical: the average US vehicle on the road is now 12.6 years old, and the parc contains roughly 280 million light vehicles, the overwhelming majority still ICE. Even as new-vehicle share tips toward electrified powertrains, the installed base that needs filters, brakes, batteries and oil changes stays ICE-dominant until at least the early 2030s. Hybrid growth extends that tail — HEVs need all the same wear items, just on a different duty cycle. AutoZone and O'Reilly sit on the same structural cushion; Genuine Parts is simply the most exposed pure-play.

The clearest expression of the transition is the diverging P&L inside the same auto makers. Ford's Q2 2026 disclosure showed ICE vehicles generating $2.62 billion of profit while the EV unit lost $722 million pre-tax, even as losses narrowed for the third consecutive quarter to $919 million adjusted. GM reported a $2.3 billion EV manufacturing-realignment charge and net income to stockholders of $1.3 billion, down 31% YoY — partially offset by ICE volume growth of 30K units. Both companies raised full-year 2026 profit guidance because ICE profitability more than covered EV losses.

Tesla's Q3 print is the cleanest signal in the data set. The company delivered 486,532 vehicles globally, beating the Bloomberg consensus of roughly 463,000 and sending the stock up ~5%. US sales, however, fell 26% YoY to ~40,816 units; Tesla's share of the US EV segment dropped to 50% in Q3 2026, the lowest on record (down from 62%). The narrative of US EV demand collapse is partly a Tesla-share story masquerading as a TAM story. Rivian sold 11,405 vehicles in Q2 2026 and remains a fraction of the segment.

Hertz provided the most direct evidence that EV residual-value math has not worked. Q2 2026 results showed Net Depreciation per Unit of $302, record Q2 revenue per day (excluding the 2022 pandemic peak), and a 1% smaller fleet producing 10% more revenue — but management is shifting back toward ICE vehicles and selling roughly 20,000–30,000 EVs, after EV repair costs and resale prices collapsed. Recall volume was up over 300% YoY in Q2, hitting an average of 15,000 vehicles per month. Hertz's experience is the cleanest realized-data test of EV resale economics available, and the verdict was negative.

Used-EV share is the trailing indicator: Cox Automotive's monitors put used-EV share at 2.4–2.8% of used-vehicle sales through summer 2026, well below new-vehicle EV share of 7.4%. The pricing gap is closing, but only because BEV new-vehicle prices are being cut — not because used-EV values are rising. Avis Budget has not disclosed fleet-mix changes at the same granularity, but the residual-value math that drove Hertz's reversal applies identically to its fleet of ~600,000 vehicles.

Short-term (months–quarters): Middle East tension keeps product inventories tight; Valero and Marathon Petroleum continue to print. ICE-aftermarket growth at Genuine Parts and AutoZone holds at low-single digits. Tesla defends its delivery line via mix and energy-storage deployments (13.7 GWh in Q3). Ford and GM keep using ICE cash flow to fund EV rationalization, with EV losses narrowing.

Long-term (1–3 years): IEA's demand cuts telegraph 2026 as a peak oil-demand year. US electrified share moves from 27% toward 35–40% by 2028, with the marginal share loss to BEV. Refining capacity utilization becomes a stranded-asset question once a recession overlaps the EV step-up. Tesla's US EV-segment share below 50% is more likely a permanent reset than a cyclical trough. The aftermarket cushion at Genuine Parts and AutoZone persists, but the secular clock is ticking on its duration.